Serhiy O. Semerikov, Tetiana A. Vakaliuk, Iryna S. Mintii, Vita A. Hamaniuk, Vladimir N. Soloviev, Olga V. Bondarenko, Pavlo P. Nechypurenko, Svitlana V. Shokaliuk, Natalia V. Moiseienko, and Vitalii R. Ruban
The article provides an overview of the origins and current state of machine vision systems, examples of machine vision problems. The article describes the use of computer vision systems in education in both conventional and pandemic conditions. The COVID-19 pandemic has triggered changes in education that have modified existing educational applications of computer vision systems and spawned new ones, including social distancing, facial mask recognition, detection of infiltration into universities and schools, prevention of vandalism and detection of suspicious objects, attendance monitoring, recognition of emotions on faces in and without masks. Computer vision systems can also be used in education to introduce immersive educational resources. On the basis of the analysis of autonomous libraries for the identification of dynamic objects, it is concluded that in the creation of computer vision systems for educational purposes it is advisable to use computer vision libraries based on in-depth learning (in particular, the implementation of convolutional neural networks). A prototype computer vision system developed on the basis of Microsoft Cognitive Toolkit and deployed in the Microsoft Azure cloud is described. The system allows you to perform with a high degree of reliability the main functions: identification of emotions and the presence of a mask on the face, as well as allows you to determine sex, age, hair color, smile intensity, the presence of makeup, glasses, etc.